governed-compression

Public research implementation surface for governed vector and KV-cache compression.

Learn more at hummbl.io.

Purpose

This repo exists to provide a better implementation surface than the current fragmented TurboQuant-adjacent landscape.

Core goals:

• CPU reference implementation first
• reproducible benchmarks
• Windows via WSL2 first
• tuple-based experiment logging from day one
• method comparison across TurboQuant-style, QJL, and simple baselines

Initial Scope

• vector encode / decode
• approximate dot product
• distortion metrics
• simple benchmark harness
• experiment logging

This repo is not yet a full inference-runtime integration project.

Layout

governed_compression/
  core/
  bench/
  logging/
tests/
examples/
docs/

Quick Start

python3 -m venv .venv
source .venv/bin/activate
pip install -e .
python -m governed_compression.cli

Repository Health

See REPO_HEALTH.md for the authoritative repository health contract and validation checks.

Current Status

Stage 1 scaffold only.